Acknowledgment The authors thank L. D. Whittig and A. Hardan of this department for the x-ray powder diffraction patterns; E. H. Stanford of the Agronomy Department for the alfalfa samples; and H. H. Sephton, National Instilutes of Health, Bethesda, Md., for thr samples of Dglycero-D-manno-octulose and D-glycero-Dido-octulose.
Literature Cited (1) Bevenue. A . , White. L. M.. Secor, G. E., Williams. K. T., J . Assoc. O$ic. Agr. Chemists44, 265 (1961). (2) Bevenue, A., Williams, K. T . , Arch. Biochem. Biophys. 34, 2.25 (1951). (3) Bland, B. F.. Dent, J. W., J . Brit. GrasslandSoc. 17, 157 (1962). (4) Bonner, J.. “Plant Biochemistry,” JJ. 25, Academic Press. New York, 19.50. (5) Caputto, R., Leloir, L. F . , Cardini, C. E., Paladini, A. C.. J . Biol. Chem. 184, 333 (1950).
(6) Charlson, A. J., Richtmyer, N. K., J . A m . Chem. SOC. 81, 1512 (1959). ( 7 ) Ibid., 82,3428 (1960). (8) Dische, Z., J . Biol. Chem. 204, 983 (1953). (9) Harper, -4. E., Elvehjem, C. A., J. AGR.FOOD CHEM.5,754 (1957). (10) Hirst, E. L.. Mackenzie, D. J.. Wylam, C. B., J . Sci. Food Agr. 10, 19 (1959). (11) LaForge, F. B., J . Biol. Chem. 28, 511 (1917). (12) LaForge, F. B., Hudson, C . S., Ibid., 30,61 (1917). (13) Leloir. L. F., Cardini, C. E., Ann. RPU.Biochem. 22, 179 (1953). (14) McComb, E. A , Rendig, V. V.. Arch. Biochem. Biophys. 99, 192 (1962). (15) McComb, E. A., Rendig, V. V., .Tatwe 193, 1101 (1962). (16) McCready, R. M . , McComb, E. A., Anal. Chem. 26, 1645 (1954). (17) Mehrad, B.. iM. S. Thesis, Universitv of California, Davis, 1964. (18) Myhre. D. V., Smith, F., J. AGR. FOOD CHEM.8,359 (1960). (19) Nordal. L4..Benson, A. A,, J . A m . Chem. Sot. 76, 5054 (1 954).
(20) Percival, E. G. V., Brit. J . :Vulr. 6, 104 (1952). (21) Plice, M. J., J . Range .Management 5 . 69 (1952). (22j Rendig,’V. V., McCornb, E. A., Plant Soil 14, 176 (1961). (23) Rendig, V. V., McComb, E. A., Arch. Biochem. Biophys. 96, 455 (1961). (24) Ibid., 99, 409 (1962). (25) Robison, R., MacFarlane, M. G., Tazelaar, A., Nature 142, 114 (1938). (26) Trevelyan, W. E., Procter, D. P., Harrison, J. S.,Ibid., 166, 444 (1950). (27) Webber, J. M., .Idv. Carbohjdrate Chem. 17, (1962). (28) White, L. M . , Secor, G. E., Anal. Chem. 33,1287 (1961). (29) Williams, K. T., Bevenue, A., J . Assoc. Offic. Aer. Chemists 36, 969 (1953). (30) Williams. K. T., Potter, E. F., Bevenue, A., Ibid., 35,483 (1952).
-
Received for review .Vouenibrr 1.1> 7963. Accepted February 6, 796.1. This inmstigation was sup@orted in part by a grant (RG-7774) from the .l‘ational Institutu of HPalth, Bethesda, Md.
FERTILIZER TECHNOLOGY
Granulation Characteristics of a 5-4- 12 (5-10-15) Fertilizer Containing Potassium Nitrate
D. R. BOYLAN and D. V. KAMAT Engineering Experiment Station, Iowa State University, Ames, Iowa
Factors influencing the granulation of a mixed fertilizer containing potassium nitrate have been studied. Kilogram batches of a 5-4-1 2 (5-1 0-1 5) mixed fertilizer containing potassium nitrate were granulated at different feed moisture levels and at inlet gas temperatures ranging from 405” to 575” F. Optimum conditions of moisture in the feed and inlet gas temperature were determined for maximum yield of the granular product and uniform distribution of the three major nutrient elements (N, P, K) in the various product size fraction!;. Similar experiments were performed with a 2-4-1 2 (2-1 0-1 5) mixed fertilizer containing potassium chloride. This grade represents a simple substitution of KCI for KN03 so physical changes due to KN03 could be more closely compared. A comparative study has been made of the results obtained with the two mixtures. The experiments were based on a statistical central composite rotable design, and the results were statistically analyzed. An attempt has been made to relate mathematically the yield and total absolute deviation in nutrient analysis between product and feed with moisture in a feed and temperature of granulation.
P
OTASSIUli nitrate has a high agronomic value as a source of both nitrogen and potash. but high cost has been the chief reason for its limited use as fertilizer. The new production facilities of Southwest Potash Corp. for fertilizergrade potassium nitrate, 13% S? 36.5Y0 K (44% K D ) . may make this material economically attractive. ‘Thishas stimulated considerable interest in the properties of potassium nitrate with respect to its behavior in mixed fertilizers during processing and its effect on the physical quality of the resulting products.
Previous Work During the past few years, much work has been done on the granulation char-
acteristics of mixed fertilizers. Pilot plant studies conducted at T V h ( 3 ) shoived that in production of very Ioxv nitrogen grades [e.g., 3-3-9 (3-12-12) or 4-6-13 (4-16-16)] the amount of moisture required for granulation of the charge material was high (14 to 167,): and the products were rather wet. Smith ( d ) . in the study of temperature and moisture relationships in granulation, noted that the utilization of ammonia as anhydrous or nitrogen solutions reduced not only the cost of nitrogen but also the free kvater content. A high salt solution phase contributed to rapid crystallization and therefore aided formulation of granules during agglomeration. These grades contained potassium nitrate as the source of potassium. VOL.
Recently, Hardestv and his coworkers, working with grades containing potassium nitrate, noted that slightly less moisture was required for agglomeration with potassium nitrate mixtures (?), but they produced a “popped-corn” shape of granules which effloresced on drying. Comparison of the yield and homogeneity of the product with mixtures that contained potassium nitrate and potassium chloride was not reported.
Experimental Work Experiments were performed to study rhe effect of moisture in the feed and inlet gas temperature on the yield and the distribution of plant nutrients in the various product size fractions. Two
1 2 , NO. 5 , 5 E P T . - O C T .
1964
423
/
Upper secfion
-(
5" -.
-. " 32 "".' thickness stainless Figure 1. sembled
steel pipe
PhotopraDh of the ammoniator-mixer
disor,Paddle
grades were prepared-54-12 (5-1015) tobacco fertilizer with potassium nitrate as the major source of potassium, and 24-12 (2-10-15) mixed fertilizer with potassium chloride as the major source of potassium (Table I). The solid raw materials (all crushed to -20 mesh) were mixed in a batch reactor, and then reacted with sulfuric acid and ammonia. The required moisture content was obtained by addition of predetermined amounts of water to the feed mixtures. The batch reactor was constructed from a section of 5-inch o.d., type 304, stainless steel tube with a wall thickness of inch (see Figures 1 and 2). A stainless steel stirrer, with a hallow shaft and four paddles, was used to mix the ingredients and to distribute ammonia. The feed materials were screened and stored in individual bins. As needed, the required quantities were removed from the bins and dry mixed in the mixer. The water, sulfuric acid, and anhydrous ammonia were added one a t a time to the mixture and mixing was continued Before for a predetermined time. transferring the mixture to the granulator dryer, duplicate samples were removed
Figure
2.
Sectional drawing of ammoniator-mixet
3). The dryer speed was 7.4 r.p.m. The product was dried, screened, and analyzed far nitrogen, phosphorus, and potassium. Experiments on both grades were statistically designed (1). The two major factors, per cent moisture in the feed and average inlrt gas temperature, were varied over a reasonable range. Preliminary experiments showed that even a variation of 0.5% moisture in the feed affected the yield of granules appreciably with the 5-4-12 (5-10-15) grade, while 1.5% moisture variation was necessary to get an approximately equivalent effect on the yield of granules with the 2-4-12 (2-10-15) mixture. The experiments with the 54-12 (5-10-15) grade were
Two mathematical models were set up-model I, relating the per cent yield with per cent moisture in the feed and temperature of the inlet gas, and model 11, relating the total absolute deviation of the three nutients N, P, and K in the product from that in feed with per cent moisture in feed and temperature of the inlet gas. The total absolute deviation was the statistic which expressed the variation of the plant nutrients between the feed arid the granular product. I t was estimated for every experimental run by summing up the individual absolute differences far N, P, and K between feed and -G 20 granular product.
+
Y
=
ba f b m
+ bzxz + b
d f b i d
1405, 6.5Ii
I
1491, 6.51
I
I
c434.61
1575,
I
-.
1405, 16)
65)
I
1
Temp. 'F
Temp' 'F 1430, 1451
1556.6)
1490, 5.79)
1550, 14.51 (49913.82)
I Figure 3. Photograph of granulatordryer and granulated product
Figure 4. Schematic o f central composite rotatable design for the 5-4-1 2 (5-1 0-1 5) grade containing potassium nitrate
15
14
ar nperature on the yield o f - 6 + 2 0 prooucr size ~ r o c r i o nTor both fertilizer grades a t their central moisture levels
6 0.51 ,
Figure 5. Schematic of central composite rotatable design for the 2-4-1 2 (2-1 0-1 5) grade containing potossium chloride
17
16
18
I
0
7.5
55
Figure 7. Effect o f moisture on the yield of -6+20 product size fraction for both fertilizer grader at their central temperotui
c
I
I
.o 0.5
,-2-4- 12(2-10-15) grade
:: 0.2
r)
r)
0
0
0.1 c
c " 0
c " 0
400
500
600
Temperature OF (inlet gas) Figure 8. Effect of temperature on the total absolute deviotion for both fertilizer grades at their central moisture levels
5.0
5.5
6.0
6.5
25
20
% Moisture in feed (day basis)
Figure 9. Effect of moisture on the total absolute deviation for both fertilizer grades ot their central temperature level V O L . 1 2 , NO. 5 , SEPT.-OCT.
1964
425
Model I :
and V. Tables I1 and I11 show the effect of moisture and temperature on the yield of granules, and Tables IV and V present the effect of moisture and temperature on the total absolute deviation in the product. The statistical analysis of these results gave the following quadratic relationships :
'l
=
45.40
+ 3.36
[ G I +
(5-10-13) grade.
with the 5-4-12
[%e] [T] [%;$I2 [T] [q [e] ] 3.05
13.37
-
8.04
160
-
XZ
- 490
XI
4.46
-
-
-
1.92
with the 2-4-1 2 (2-1 0-1 5) grade.
x/
Optimum yield 4 6 . 3 % a t 5 0 7 O F and 6 . 5 5 % rn oisture w i t h 5-4-12(5-10-15)
Model 11: I-
=
0.0960
+ 0.0175
0.0125
CT]
0.0664
[T]
0.0464
X?
XI
["I
~
,,"""I
- 6.5 - 490
t
+ '+
[-I2 [7 1 [05-1 -
-
0.0575
XI
- 490
XP
- 63
Table II. Effect of Moisture and Temperature on the Yield of Product for the 5-4-12 (5-10-15) Grade Containing Potassium Nitrate Drying time was 20 minutes. Moisture in the product varied from 0.38yc to 0.927'. Average product temperature varied from 203' to 256' F. Figure 10. Response surface showing contours of constant yield for the 5-4-1 2 (5-1 0-1 5) grade containing potassium nitrate
Yo Yield
I Optimum y i e l d 43.8'10 4 7 7 O F and 15.8% moisture w i t h 2-4-12 (2-10-15)
/
1
I
Run
Temp. of Gas,
Table 111. Effect of Moisture and Temperature on the Yield of Product for the 2-4-12 (2-10-15) Grade Containing Potassium Chloride Drying time was 45 minutes. Moisture in the product varied from 0.297, to 0.92%. Average product temperature varied from 194' to 249' F. Temp. of
Response surface showing contours of constant yield for the 2-4-1 2
(2-1 0-1 5) grade containing potash chloride
426
AGRICULTURAL A N D F O O D CHEMISTRY
Yield
-6+20,
No. F. % % 13.8 5.75 492 K-10 27.8 6.99 552 K-11 27.1 6.98 432 K-12 23.1 7.23 K-13 31 .O 5.98 K-16 22.6 6.01 K-18 36.2 6.50 K-2 1 33.4 6.51 K-27 45.4 6.50 490 K-29" a K-29 readings were estimated by averaging the 5 replications conducted at the central point (Figure 4).
Run No.
Figure 1 1.
Moisture in Feed,
Gor,
Moisture in Feed,
Yield
-6f20,
F. % % 16.05 41.4 575 K-32 15.87 30.9 406 K-33 18.15 33.9 495 K-34 13.94 24.1 491 K-35 17.49 32.1 552 K-36 14.62 27.2 553 K-38 14.53 25.8 430 K-42 17.54 31.5 432 K-46 16.08 43.4 494 K-470 a K-47 readings were estimated by averaging the 5 replications conducted at the central point (Figure 5 ) .
Table IV.
Moiatun, in Feed,
Temp. of
R""
Gor.
NO.
OF.
Effect of Moisture and Temperature on Total Absolute Deviation of Plant Nuti for 5-4-12 (5-10-15) Grade Containing Potassium Nitrate
K-10 492 K-11 552 K-12 432 K-13 492 5sn K-16 ~~. K-18 627 K-21 574 K-27 406 K-296 470 a Total. * K-27 readings were
N
5.75 6.99 6.98 7.23 5.78 6.01 6.50 6.51 6.50
5.31 5.45 5.79 5.57 5.38 5.51 5.23 5.60 5.53
Temp. o f GO$,
Moirture in Feed,
NO.
OF.
%
N
16.05 15.87 18.15 13.94 17.47 16.62 16.53 17.54 16.08
2.26 2.48 2.32 2.41 2.36 2.37 2.39 2.47 2.45
K-32 575 K-33 406 K-34 495 K-35 491 K-36 552 K-38 553 K-42 430 K-46 432 K-47s 494 0 Total. * K-47 readings were
=
the o.1900
5-4-12
0.0380
XI
0.0063 o'0200
lK201
12.58 12.77 12.65 12.44 12.69 12.65 12.67 12.58 12.68
(15.15) (15.40) (15.25) (15.00 (15.301 (15.25) (15.30) (15.15) (15.27)
N
0.08 0.10
0.15 0.11
0.14 0.06 0.12 0.14 0.02
P 0.01
0.04 0.05 0.03 0.05
TOtOi
K
Ab,. De".
0.08 0.08 0.08 0.04
0.17 0.22 0.28 0.18 0.31 0.14 0.21
0.10
0.08 0.04 0.04 0.06
0.00
0.05 0.01 0.01
0.19 0.09
of
-6
P'
5.83 5.82 5.74 5.91 5.96 5.84 5.86 5.87 5.92
Abr. De". of Nutrients Between -6 20 Size Product and Feed N P K
+
+ 20 Size Product
(PlOSI"
K
(K+OI
(13.38) (13.37) (13.62) (13.51) (13.63) (13.37) (13.41) (13.46) (13.57)
12.81 13.23 12.92 12.98 12.92 12.81 13.19 12.72 12.79
(15.50) (15.85) (15.60) (15.65) (15.60) (15.50) (15.80) (15.60) (15.45)
0.21
0.03
Total
Abr. Der.
0 .12
.30 .04 .17
0.24
0.03
0.09
0.02 0.02 0.02
0.07 0.05
.12 0.00 0.12 0.17 0.12
0.36 0.34 0.07 0.29 0.22 0.27 0.23 0.26 0.17
estimated by averaging the five replications conducted at the central point (Figure 5).
xX -
grade.
1
470
[ 60 - 16.0 ~
x2
- 160
+
-
X,
- 490
-
[ -T I x1
16.0
[T]
Figure 12. A:
Nutrient A n d y &
- 490 +
0.0737
(13.57) (13.71) (13.58) (13.64) (13.74) (13.43) (13.52) (13.69) (13.48)
~
(5-10-15)
+ 0.0038
+
K
Effect of Moisture and Temperature on Total Absolute Deviation of Plant Nutrients for 2-4-12 (2-10-15) Grade Containina Potassium Chloride
R""
with
5.72 5.98 5.92 5.95 6.00 5.86 5.70 5.92 5.87
Abr. De". of Nutrientr betwe 20 Size Product and k e a
+ 20 S i e Product
estimated by averaging the five replications conducted at the central point (Figure 4).
Table V.
Y
%
Nutrient Analysis of - 6 Pa (PeOsI"
with the 2-4-12 (2-10-15) grade. The effect of temperature on the yield with the two mixtures may he shown graphically (Figure 6 ) . T h e yield was generally higher with the 54-12 (5-1015) formulation. Because the tendency to form large lumps was greater a t higher temperatures with the 24-12 (2-10-15) mixture, the yield decreased more rapidly. Moisture was more critical in the 5-4-12 (5-10-15) than in the 24-12
Photographs of granular products
-6+8 size fraction of the 5-4-12 (5-10-151 grade containing potarsiurn nitrote .. 8: -8f16 size fraction of the 5-4-12 15-10-151 grade containing potoiiium nitrate C: -6+8 size fraction of the 2-4-12 (2-10-151 grade containing potaiilvm chloride D: -81-16 size fraction of the 2-4-12 (2-10-151 grade containing mtaisium chloride
(2-10-15), and affected the yield to a greater extent (Figure 7). The effect of temperature a n the total absolute deviation with the two mixtures may be shown graphically (Figure 8). T h e total absolute deviation was much lower with the 54-12 (5-10-15) mixture than with the 24-12 (2-10-15) mixture, showing the product to be more homogeneous with mixtures containing potassium nitrate. With the increase of moisture in the feed, the total absolutedeviationdecreased and, after attaining a minimum, started increasing with the 5 4 - 1 2 (5-10-15) mixture. With the 2 4 1 2 (2-10-15) it continued to decrease; but was considerably higher at the optimum yield condition than with the 54-12 (5-1015) mixture (Figure 9). Three-dimensional views showing the effect of temperature and moisture on the yield of product are shown in Figures 10 and 11.' T h e lines show ConLours of constant yield. T h e distance between the adjacent contour lines parallel to temperature axis is greater for the 2 4 1 2 (2-10-15) mixture than for 54-12 (5-10-1 5) mixture, indicating that moisture is more critical with mixtures containing nitrate of potash. T h e optimum yield of 46.3y0 was noted a t 507' F. and 6.55% rnoisturewith the 54-12 (5-10-15) mixture, and the optimum yield of 43.XY0 a t 447' F. and 15.X% moisture with the 2-4-12 (2-1015) formulation, T h e granules obtained with the 5-4-12 (5-10-15) mixed fertilizer containing
V O L . 1 2 , NO. 5, SEPT.-OCT.
1964
427
potassium nitrate were more spherical and uniform1)- shaped than those obtained lvith the 2-4-12 (2-10-15) mixed fertilizer containing potassium chloride (Figure 12). The 5-4-12 (510-15) mixed fertilizer also gave more cohesive, less porous, and tougher granules. The granule product in both mixtures {vas relatively more homogeneous than
Olive, B. i f . , Farm. Chem. 124, No. 2, 32 (1961). (3) Hein, L. B., Hick, G. C., Silverberg, J., Searz. L. F., J. AGR. FOODCHEM. 4, 318 (1956). (4) Smith, R. C., .4gr. Chem. 11, No. 2, 34 11956).
reported in the literature since the raw materials used in this study were of finer mesh size.
(l)
‘Ochran>
’‘*
G’,
‘Ox,
G’
Receioed f o r review .Vorlember 1.5, 7963. Accepted .March 16, 1964. Division of Fertiiirer and Soil Chemirtry, 145th .Meeting, ACS, .Vew York, September 1963.
Designs,” 2nd ed.: Wiley, New York, 1960. (2) Hardesty, J. 0.: Dickey, G. F.,
FERTILIZER TECHNOLOGY
A One-Step Continuous Quick-Curing Triple Superphosphate Process Employing Rod Mill Grinding
I
D. R. BOYLAN and A. B. AMlN Engineering Experiment Station, Iowa State University, Ames, Iowa
An investigation was undertaken to determine the feasibility of a one-step quick-curing continuous process for the manufacture of triple superphosphate. Bench-scale work was carried out in a 1-quart, laboratory ball mill made of stainless steel. The results from this preliminary work indicated that quick-curing triple superphosphate of high conversion could be produced. The process was investigated on pilot-plant scale with successful operation. Completely cured powdered triple superphosphate was obtained within 1 hour.
T
a quick-curing continuous superphosphate process include the following: the product can be shipped directly, thus reducing the required storage facilities and process inventory; operating conditions in the mixing step can be chosen with more flexibility; control of the process can be essentially automatic; uncertainty as to production rates can be eliminated to some extent because a finished product can be produced in a matter of hours, instead of weeks as required by the storage curing process; and there is a possibility of lower operating labor and maintenance requirements and elimination of a final crushing and screening step since finely ground products can be obtained directly.. HE ADVASTAGES O f
Previous Work
Many processes (3,5, 9, 72) have been tried for quick-curing of triple superphosphate. None of these processes has been able to eliminate curing time entirely without loss of phosphorus availability. In some processes, both drying and storage curing are employed to shorten the period of curing. A quick-curing process is therefore desirable in which a minimum amount of phosphoric acid is needed to convert a maximum amount of rock into available form without curing. The present paper
428
discusses an investigation of such a process. The development of a quick-curing, one-step process for normal superphosphate was carried out by Rounsley and Boylan (74) and Martinez (8),by grinding and drying in a single piece of equipment. This process was extended to the manufacture of triple superphosphate. Essentially the same laboratory and pilot plant equipment were used. Process Variables
The important process variables which effect the reaction stage and curing stage of the process are acid concentration, temperature, acidulation ratio (acidulation ratio as used here is the weight ratio of acid P205 to rock PnOa), and time. The effects of these variables using grinding in a laboratory ball mill and a pilot plant rod mill were investigated. The effect of acid concentration is the most important process variable. An acid concentration of about 70% H3P04 has been found most suitable, considering the physical properties of the mixture during the process and the degree of completion of reaction. Higher acid concentrations result in poor conversion, probably due to excessive side reactions. Very low acid concentrations result in
AGRICULTURAL A N D FOOD CHEMISTRY
slow reaction and incomplete curing with a resulting product of high moisture and free acid content. Temperature is also an important process variable. Its effect interferes with acid concentration because the reaction is exothermic. Acid concentrations above 75% H3P04 increase the rate of reaction, and an excessive amount of heat is liberated within a short time in the first stage of the reaction. Temperatures above 284’ F. may decompose the monocalcium phosphate into unavailable pyrophosphate ( 7 7 ) . Rapid rise in temperature of the freshly acidulated mass tends to drive off water, increasing the concentration of the acid. The increased concentration means further increase in reaction rate, and hence higher temperature of the mixture. Thus a balance is required between the temperature and the acid concentration such that optimum conditions are obtained in the first stage of the reaction. Various workers (2-4, 6, 75) have suggested limiting temperatures in the range ofl5O’ to302’F. Acidulation ratio is less important as a process variable than as an economic factor. An acidulation ratio of 2.0 is the theoretical minimum based on stoichiometry. Generally, it is necessary to use an acidulation ratio greater than theoretical due to the loss of acid consumed by impurities and side reactions.
